System identification of underwater vehicles

Abstract

The aim of this project was to obtain hydrodynamic derivative measurements from sea trials using an underwater vehicle which is a half-scale model of the PAP-104 mine countermeasures UUV. Trials data were collected for a series of manoeuvres which were then reconstructed to give the required vehicle state estimates. This reconstruction was carried out using an extended Kalman filter. Hydrodynamic derivative estimates were then generated by a stepwise regression algorithm, which took the reconstructed vehicle state and measured vehicle forces as input. Traditionally, these types of hydrodynamic measurements are made using a planar motion mechanism (PMM). Ideally system identification (SI) techniques would provide a more practical alternative, as the UUV need only be driven in a predefined way and the SI techniques applied in order to obtain current hydrodynamic measurements. SI techniques also give measurements based on the overall response of the vehicle and umbilical, which is not the case with PMM measurements.